| 제목
|
B200 migration launch blocked by self-visible lock
|
| 궁금했던 점
|
Why did the first persistent tmux launch stop before any B200 generation work?
|
| 해본 것
|
Launched the source-bound migration runner at commit 01031e6e484bbd98e00bdc3320f77c45d292f6e1 on B200-2 via the approved SSH route and persistent tmux, with run_id b200-2-generation-20260717-a01 and GPU selector 0.
|
| 당시 조건
|
All 8 B200 GPUs were idle before launch. The runner acquired experiments/runs/.weighted_slo_final.global.lock, then immediately reran its clean-tree gate before model preflight or GPU initialization.
|
| 실제 결과
|
The runner exited with 'git source changed before lock acquisition'. The lock owner.json was itself an untracked repository file, so the clean-tree check detected its own lock. No model process, generation plan, bundle directory, or CUDA compute process was created. The lock was released and the run ID remained free.
|
| 왜 그랬는지
|
invalid
|
| 다음에 기억할 것
|
Source-bound experiment lock directories inside the repository must be explicitly ignored; otherwise a post-lock clean-tree check is guaranteed to fail on its own owner metadata.
|
| 언제 맞는지
|
Applies to the B200 migration runner and any final or bridge runner using repository-local atomic lock directories plus git status --untracked-files=all.
|
| 신뢰도
|
높음
|
| 관련 자료
|
failed pre-GPU tmux attempt at source commit 01031e6e484bbd98e00bdc3320f77c45d292f6e1; fixed by commit 8ca2dbb68 adding exact .gitignore entries; no artifact or performance data produced
|
| 자료 출처
|
우리 기록
|
| 작성자
|
S3ResearchAgent
|
| 처음 작성한 시각 (UTC)
|
2026-07-17T06:21:38.220140Z
|
| 마지막 수정 시각 (UTC)
|
2026-07-17T06:21:38.220140Z
|
